Pp. 366 and 385. Work of Deformation up to Break-
ing. — This may be very closely estimated without the necessity
of drawing out the stress-strain diagram, using an approximate
formula devised by Kennedy. Imagine the diagram divided by
a horizontal line passing through the elastic limit, and let r repre-
sent the ratio of limit to maximum load, while \V is the maximum
and total load. Then the lower portion is nearly a rectangle of
area r WA. The upper portion may be taken as a parabola whose
area s w—
.-. Total work - rW A +• - (W- r\y) A = WA
>j \ '
P. 369. Coefficient of Linear Expansion. — Invar is
the name given by its discoverer, M. Guillaume, to an almost
non-expansible material, consisting of Steel alloyed with 36% of
Nickel, Its coefficient of linear expansion has the value '000,00087
or 87 x io~8. This remarkable property of low expansion makes
the metal useful for many pieces of physical apparatus, such as
clock pendulums, &c. Neither is the expansion anything so im-
portant as the coefficient would indicate, for the full alteration in
length, as indicated by this value, does not accrue in less than
several days, and the real expansion for change of temperature is
therefore very small indeed, and practically negligible.
P. 3J2. Testing Machines. — A very fine example of a
Wicksteed machine has been built for the French Government.
It is a legitimate descendant of the Wicksteed-Kennedy machine
on p. 371, and can be best understood by a previous reading of
p. 372. It is illustrated in Fig. 950, and has 300 tons load
capacity, while admitting specimens 88 ft. long and 3 ft. 3 ins.
diameter. In fact, it may be at once explained that the machine
is not merely to demonstrate, or to test small specimens repre-
senting the material of structures, but is for considerable portions
of the structures themselves, such as built-up pillars or girders,
riveted joints, &c. It is hoped that by its aid much information